Week 10 — Reporting results clearly

Scope & method. Rapid narrative review, not a systematic review. Structured search of the sport-science and physiotherapy reporting literature (PubMed, BMJ Open, BJSM). Findings reported as the cited studies state them; the reporting guidelines themselves (CONSORT, STROBE) are cited as standards.

The principle under review Week 10 · completeness and calibration

Week 10 argues that reporting should be complete (all preregistered outcomes, including nulls) and calibrated (language, emphasis and causal claims matched to the evidence), and that reporting-guideline checklists such as CONSORT and STROBE help achieve both. The empirical questions are whether the field reports the necessary methodological detail, and whether abstracts and conclusions are free of “spin” — reporting strategies that make an effect look larger or more definite than the data support.

1 · How well is it currently applied? the evidence

Spin is highly prevalent

Where spin has been measured directly, it is common. Kumar, Prakash, Hariohm and Naveen Kumar (2023), examining physiotherapy RCTs with statistically non-significant results, found spin in 90% of abstracts and 88% of main-text conclusions — that is, the great majority of “negative” trials were written to imply a benefit the data did not support. This matters because experimental work shows spin changes readers' conclusions: physiotherapists reading a spun abstract rate the intervention as more beneficial than the same results reported plainly.

Key methodological details are under-reported

Completeness of methods reporting is poor against CONSORT items. Schulz and colleagues (2022), reviewing 163 sports-medicine and orthopaedic trials, found that only 4% reported the blinding status of all stakeholders, 18% reported the randomisation type, method and allocation concealment, and 60% reported a sample-size calculation. Under-reporting of allocation concealment, blinding and randomisation implementation is exactly the pattern found in reviews of CONSORT adherence in sports-medicine journals, and it prevents readers from judging a trial's risk of bias.

The record is skewed toward positive framing

Selective and positive reporting compounds the problem. Büttner and colleagues (2020) found 82% of sport and exercise medicine studies reported supported hypotheses, and Twomey and colleagues (2021) an ~81% positive-result rate — figures that, together with the spin data, indicate a literature in which negative and null findings are both under-reported and, when reported, framed positively. This biases the evidence base toward apparent effects (Weeks 1, 4).

Source	Finding	Implication
Kumar, Prakash, Hariohm & Naveen Kumar (2023) Physiother. Pract. Res.	Spin in 90% of abstracts and 88% of conclusions of non-significant physiotherapy RCTs	Negative results are routinely framed as positive
Schulz et al. (2022) BMJ Open	4% reported blinding of all stakeholders; 18% randomisation/allocation concealment; 60% sample-size calculation	Key CONSORT items are severely under-reported
Büttner et al. (2020) Br. J. Sports Med.	82% of SEM studies reported supported hypotheses	Selective/positive reporting is the norm
Twomey et al. (2021) Comms. in Kinesiology	~81% positive-result rate; limited registration and data	Nulls are under-represented in the record
Schulz, Altman & Moher (2010) BMJ (guideline)	CONSORT: the standard for complete RCT reporting	The standard exists to prevent these gaps
von Elm et al. (2007) Ann. Intern. Med. (guideline)	STROBE: the standard for observational reporting	Applies to the field's many non-randomised studies

2 · Is there cause for concern? objective assessment

Reporting is weak on both dimensions the week identifies. Completeness is poor — core CONSORT items such as blinding and allocation concealment are reported in a small minority of trials, and nulls are under-represented. Calibration is worse — spin appears in around nine in ten non-significant physiotherapy trial abstracts, and the literature is dominated by positive framing. Because reporting is the interface between a study and the people who act on it, these distortions propagate directly into practice and into biased meta-analyses.

The concern is offset by the availability and increasing adoption of remedies. Reporting-guideline checklists (CONSORT, STROBE, and the EQUATOR network) are established and increasingly required by journals, spin is now being measured and named in the field, and preregistration provides a reference against which selective reporting can be detected. The tools to close the gap exist; the shortfall is in their consistent use and enforcement.

3 · Immediate steps to improve actionable

1. Report every preregistered outcome, including nulls, and cross-check the paper against the preregistration to prevent selective outcome reporting.
2. Work the relevant checklist as you write — CONSORT for trials, STROBE for observational studies — so blinding, allocation, randomisation and flow are all reported.
3. Make effect sizes with confidence intervals the centrepiece, interpreted against a smallest effect of interest, rather than bare p-values.
4. Run a spin self-test: read the abstract and conclusion alone and ask whether a reader would infer something stronger than the data support; if so, remove it.
5. Match adjectives and causal verbs to the evidence — call a small effect “small,” and reserve causal language for randomised designs (Week 11).

References